Phase I Study of Trifluridine/Tipiracil Plus Irinotecan and Bevacizumab in Advanced Gastrointestinal Tumors

This article is featured in Highlights of This Issue, p. 1533

Fluoropyrimidines [5-fluorouracil (5-FU) or capecitabine] form the backbone of active combination regimens for metastatic colorectal cancer (mCRC) and are often combined with oxaliplatin, irinotecan, or both (1, 2). The further addition of targeted agents, such as the anti-VEGF antibody bevacizumab to these regimens has been shown to improve overall survival (OS) in the first- and second-line settings (3–5). Despite these advances, many patients develop treatment-refractory mCRC, for which there are few treatment options.

Trifluridine and tipiracil [FTD/TPI (TAS-102)] is an orally administered cytotoxic agent consisting of trifluridine (FTD), a thymidine analog, and tipiracil (TPI), a thymidine phosphorylase inhibitor (6). FTD/TPI has a unique mechanism of action that distinguishes it from fluoropyrimidines: FTD is incorporated into DNA, causing DNA dysfunction, and TPI increases FTD bioavailability (7, 8). In a phase III trial evaluating FTD/TPI versus placebo in patients with mCRC refractory to standard therapies, FTD/TPI achieved a statistically significant improvement in OS versus placebo [hazard ratio (HR), 0.68; 95% confidence interval (CI), 0.58–0.81; P < 0.001; median OS, 7.1 vs 5.3 months, respectively] and manageable toxicity (9). On the basis of these results, FTD/TPI was approved at 35 mg/m²/dose for patients with previously treated mCRC. FTD/TPI also recently received approval for the treatment of metastatic gastric or gastroesophageal junction cancer after two previous lines of chemotherapy (10).

In preclinical studies involving human colorectal cancer xenograft models, the antitumor activity of FTD/TPI combined with irinotecan or bevacizumab was significantly greater than that of each agent alone (11, 12). Consistent with these preclinical data, preliminary antitumor activity was observed with the combination of FTD/TPI and irinotecan in patients with previously treated mCRC in an initial phase I study; however, a high rate of grade ≥3 hematologic toxicities was observed, indicating that further investigation was needed to optimize the FTD/TPI–irinotecan regimen (13). On the other hand, it was expected that bevacizumab and FTD/TPI would have nonoverlapping toxicities based on prior experience with fluoropyrimidines (3–5, 14).

These findings formed the basis for exploring the combination of FTD/TPI with irinotecan and bevacizumab in patients with advanced gastrointestinal tumors, including mCRC, in a two-part, phase Ib dose-escalation/expansion study (ClinicalTrials.gov identifier NCT01916447). The dose-escalation phase examined the safety of the combination of FTD/TPI and irinotecan and characterized the MTD, and the expansion phase further evaluated the safety, preliminary activity, and pharmacokinetics of the combination of FTD/TPI, irinotecan, and bevacizumab.

Patients with histologically confirmed advanced gastrointestinal tumors (malignancy of gastrointestinal origin or an adenocarcinoma of unknown primary, likely to be of gastrointestinal origin) were eligible for enrollment in the dose-escalation phase. Eligibility for the triplet combination in the expansion phase was restricted to those with histologically confirmed mCRC. Patients aged ≥18 years with an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 or 1 at baseline who were refractory to at least one line of chemotherapy for metastatic disease and had no curative treatment options were enrolled. Key exclusion criteria included recent major surgery or extended-field radiotherapy (within 4 weeks prior to starting therapy), recent anticancer therapy (within 3 weeks prior to initiating therapy), any unresolved grade ≥1 toxicities from prior therapies, partial or total gastrectomy, a serious illness or medical condition that precluded safe participation, or active central nervous system metastases. To assess the contribution of the toxicity of FTD/TPI to the FTD/TPI–irinotecan combination, patients with prior exposure to irinotecan were eligible only if they had required no dose reductions or delays in irinotecan during their previous therapy.

The study followed the Declaration of Helsinki and good clinical practice guidelines, and approval was obtained from each institution's review board or an independent ethics committee. All patients provided written, informed consent prior to enrollment.

This was a multicenter, open-label, nonrandomized, two-part phase Ib study comprising a dose-escalation phase to evaluate the safety and to determine the MTD of the combination of FTD/TPI and irinotecan in gastrointestinal cancers, followed by an expansion phase to further investigate the safety, pharmacokinetics, and preliminary activity of FTD/TPI and irinotecan at the MTD in colorectal cancer with or without bevacizumab. Patients received oral FTD/TPI twice daily on days 1–5 of a 14-day treatment cycle, intravenous irinotecan on day 1 of every cycle at the same time as FTD/TPI, and in the expansion phase, intravenous bevacizumab on day 1 of every cycle before irinotecan. Treatment was administered until RECIST-defined disease progression or clinical disease progression, unacceptable toxicity [adverse events (AEs) leading to more than three dose reductions of FTD/TPI or irinotecan or a dose delay >14 days from the scheduled start of the next cycle], patient request to withdraw treatment, pregnancy, or physician's decision to switch to another cancer therapy.

The dose-escalation phase followed a traditional 3+3 design (Fig. 1), in which patients in sequential dose-level cohorts received escalating doses of FTD/TPI (20, 25, 30, or 35 mg/m²/dose, administered twice daily) and irinotecan (120, 150, or 180 mg/m²/dose). At least three evaluable patients were treated in each cohort, and at least six patients were treated at the MTD. Escalation to subsequent dose levels occurred only after the previous dose level was determined to be safe per protocol. Intrapatient dose escalation was not allowed. The MTD was defined as the highest dose level at which less than 33% of evaluable patients treated had a dose-limiting toxicity (DLT) during cycles 1 or 2 of study drug administration.

The expansion phase was initiated once the MTD was established, and only patients with mCRC received bevacizumab. The first six patients were enrolled in parallel and administered the MTD of FTD/TPI plus irinotecan plus bevacizumab 5 mg/kg. The expansion phase included an interim safety assessment (ISA) to assess safety in the first six patients after two treatment cycles or treatment discontinuation, whichever came first. Additional enrollment of patients into the expansion cohort for treatment with the triplet combination (up to a total of 36 patients) was contingent upon the completion of the ISA and the absence of unexpected safety concerns. In the event of unexpected AEs, another safety evaluation was planned for the next six patients enrolled, and accordingly, treatment was planned to be continued with dose adjustments for bevacizumab or the doublet combination only.

The safety population included all patients who received one or more doses of the study drug. Safety assessments were performed from the first dose until 30 days after the last dose of the study drug. AEs were graded according to the NCI Common Terminology Criteria for Adverse Events version 4.03. A DLT was defined as a cycle 1 or 2 AE that met one of the following criteria: grade 4 neutropenia lasting >7 days; febrile neutropenia, grade 4 thrombocytopenia, or grade 3 thrombocytopenia with bleeding; grade ≥3 nonhematologic toxicity; grade ≥3 nausea, diarrhea, or vomiting lasting >48 hours; any study drug-related toxicity resulting in a >2-week delay in initiation of cycle 3; or any study drug-related toxicity preventing the administration of ≥80% of the planned cumulative dose for cycles 1 or 2. Dosing and dose reductions can be found in Supplementary Materials and Methods.

Response was evaluated using investigator-assessed RECIST version 1.1, with tumor assessments at baseline and every four cycles (8 weeks) for the first 24 cycles or as clinically necessary. Patients who had completed one or more cycles of study medication and had radiologic/clinical progression assessments were evaluable for efficacy.

Pharmacokinetic assessments were performed in the first six patients and in at least 12 patients, following the ISA in the expansion phase. Samples for pharmacokinetic analyses were collected on days 1–3 of cycle 1 predose (30 minutes prior to the administration of the first study drug) and postdose (at 30 minutes and 1, 1.5, 2, 4, 6, 8, 10, 24, and 48 hours after irinotecan infusion).

Assessed pharmacokinetic parameters included maximum observed plasma concentrations (C_max), area under the plasma concentration–time curve (AUC) from time 0 to the last measurable plasma concentration (AUC_0–last) or from time 0 to infinity (AUC_0–inf), time to C_max (T_max), apparent terminal phase elimination half-life (t_1/2), oral clearance (CL/F), and apparent volume of distribution (V_d/F).

Descriptive statistics were used to summarize patient characteristics, efficacy, and safety measurements. Pharmacokinetic analyses used standard noncompartmental models and calculations. Progression-free survival (PFS) was analyzed using Kaplan–Meier methodology; CIs for median PFS were generated using the Brookmeyer and Crowley method. All data summaries and listings were produced using SAS version 9.1.3 (SAS Institute).

A total of 50 patients were enrolled between September 24, 2013, and September 8, 2016. Of these, 26 patients [21 (81%) with colorectal cancer, two (8%) with gastric/gastroesophageal junction cancer, two (8%) with pancreatic/biliary cancer, and one (4%) with gastrointestinal cancer of unknown primary] were enrolled into six dose-escalation cohorts, with three to six patients per cohort, and 24 patients (all with colorectal cancer), into the expansion cohort (Fig. 1). All 50 patients received at least one dose of the study drug. At data cutoff (March 9, 2017), 42 patients (84%) had discontinued treatment, most frequently as a result of disease progression (n = 35; 70%). The eight patients (16%) for whom treatment was ongoing at the end of the study were in the expansion cohort (Supplementary Table S1).

In the overall patient population (N = 50), the median age was 55.5 years (range, 19–85 years), and the majority (74%) of patients had an ECOG PS of 1 (Table 1). Overall, 60% of all patients had received four or more prior regimens for advanced disease, and 84% had previously received irinotecan.

In the dose-escalation cohorts, a total of two DLTs were observed at the highest dose level assessed (dose level 6; FTD/TPI 30 mg/m² twice daily and irinotecan 180 mg/m²): one patient developed grade 3 fatigue in cycle 1, and another developed grade 2 neutropenia in cycle 2, which resulted in a >2-week delay in initiating cycle 3. No DLTs were observed at dose levels 1 through 5. On this basis, the MTD for the expansion phase was established as FTD/TPI 25 mg/m² twice daily plus irinotecan 180 mg/m² (dose level 5). Bevacizumab was added to this combination at the standard mCRC dose of 5 mg/kg (15).

The median duration of treatment was 4.5 months (range, 0.2–14.0) in the overall patient population (N = 50) and 6.6 months (range, 1.5–14.0) in the expansion cohort. In the expansion cohort, 23 of 24 patients (96%) received ≥80% of the planned target dose of FTD/TPI.

In the combined dose-escalation cohorts (n = 26), all patients experienced one or more AEs of any-cause and 69% of patients experienced grade ≥3 AEs (determined to be treatment-related in 46% of patients); 27% experienced serious AEs (Table 2; Supplementary Table S2; Supplementary Results). One death because of acute renal failure (not considered treatment-related) occurred in cohort 5 (FTD/TPI 25 mg/m² twice daily plus irinotecan 180 mg/m²), and another patient in cohort 6 (FTD/TPI 30 mg/m² twice daily plus irinotecan 180 mg/m²) died because of clinical disease progression.

In the expansion cohort (n = 24), all patients experienced one or more all-cause AEs, and 83% experienced grade ≥3 AEs (Table 2); serious AEs were observed in 38% of patients. Treatment-related AEs of any grade occurred in 96% of patients and grade ≥3 treatment-related AEs occurred in 67% of patients (Supplementary Table S2). AEs were predominantly hematologic or gastrointestinal-related, with the most frequent all-cause any-grade AEs being neutropenia (in 79% of patients), fatigue (58%), nausea (58%), vomiting (54%), and diarrhea (50%). The most frequently occurring grade ≥3 AEs were neutropenia (42%), leukopenia (25%), and diarrhea (12%). Most AEs were managed using dosing adjustments: seven patients (29%) had dosing interruptions and four (17%) had dosing modifications because of AEs of any-cause. One patient discontinued study treatment as a result of an AE (rectal hemorrhage) that was not treatment-related. At the time of data cutoff, no deaths were reported in the expansion cohort. Overall, no grade 5 AEs related to treatment were reported in this study.

Pharmacokinetic parameters were evaluated in 18 patients in the expansion cohort following the first dose of FTD/TPI in combination with irinotecan and bevacizumab. After a single dose, FTD and TPI reached maximum plasma concentrations at a mean of 1.22 and 2.73 hours, respectively, and declined with a t_1/2 of 2.32 and 1.85 hours (Table 3). The pharmacokinetic exposure for FTD and TPI in the triplet combination regimen (AUC_0–inf, 5,625 and 372 ng·hour/mL, respectively) was consistent with data from previous single-agent studies (16). The C_max for irinotecan was reached at a mean of 0.54 hours after dose administration and declined with a t_1/2 of 8.72 hours; AUC_0–inf was 13,054 ng·hour/mL.

No objective responses occurred in the dose-escalation phase. As the dose-escalation phase included a heterogeneous population with respect to tumor types and dose levels and bevacizumab was not included in the dosing, efficacy was not reported in further detail for these cohorts.

All 24 patients in the expansion cohort were evaluable for efficacy. Three patients (12%) had confirmed partial responses, two of whom had received previous irinotecan and one of whom was irinotecan naïve. Seventeen patients (71%) had stable disease for a median duration of 8.1 months, and 16 (67%) patients had stable disease lasting longer than 4 months. These 17 patients had received a median of four previous systemic regimens, with 16 having received previous irinotecan; one patient was irinotecan-naïve. The median PFS was 7.9 months (95% CI, 5.1–13.4 months; Fig. 2), and PFS rates at 6 and 12 months were 66% (95% CI, 42–82) and 37% (95% CI, 12–63), respectively.

The results of this phase I trial demonstrated that FTD/TPI combined with irinotecan and bevacizumab was tolerable and showed preliminary efficacy in patients with heavily pretreated mCRC. In the dose-escalation portion of the study, two DLTs were observed (grade 3 fatigue and grade 2 neutropenia), and the MTD for the doublet combination was defined as FTD/TPI 25 mg/m² (administered twice daily on days 1–5 of a 14-day cycle) plus irinotecan 180 mg/m² (administered on day 1 of a 14-day cycle) in patients with heavily pretreated gastrointestinal tumors.

In the expansion phase, bevacizumab was added at 5 mg/kg (on day 1 of the 14-day cycle) to this regimen in patients with previously treated mCRC. The safety profile of the triplet combination, with respect to the types and frequencies of AEs, was similar to that with the doublet combination; any-grade serious AEs were reported in 38% of patients in the expansion cohort and in 32% of patients in the overall population. This finding indicated that the addition of bevacizumab did not result in cumulative toxicity or in new safety concerns. The most common AEs of any-cause and grade were neutropenia, fatigue, nausea, and vomiting, and the most frequently occurring grade ≥3 AEs were hematologic (decreased neutrophil count and decreased leucocyte count). However, AEs were managed using dosing modifications, including dosing delays and interruptions, and there were no treatment-related AEs leading to the discontinuation of treatment or to treatment-related deaths. These results indicated an overall safety profile that was consistent with that of the individual agents, with no unexpected safety concerns.

The pharmacokinetic assessment of the triplet combination regimen presented no notable changes from the pharmacokinetics of single-agent FTD/TPI or irinotecan. The AUC_0–inf values for FTD and TPI in the triplet combination regimen (5,625 and 372 ng·hour/mL, respectively) were similar to those of single-agent FTD/TPI (at 25 mg/m²/dose) in patients with solid tumors (4,297 and 222 ng·hour/mL, respectively; ref. 16). In addition, the mean clearance of irinotecan in the present study (15.03 L/hour/m²) was comparable with previously published data for single-agent irinotecan (19.3 L/hour/m²; ref. 17).

Evidence of antitumor activity was noted with the triplet combination in this heavily pretreated patient population, half of whom had received four or more previous systemic regimens. Preliminary efficacy results indicated partial responses in 12% of patients, stable disease in 71%, and a median PFS of 7.9 months in the expansion phase. For reference, in phase III second-line studies conducted in patients refractory to only one regimen for mCRC, FOLFIRI (leucovorin, fluorouracil, and irinotecan) plus ramucirumab resulted in an objective response rate (ORR) of 13.4% and median PFS of 5.7 months, whereas FOLFIRI plus aflibercept resulted in an ORR of 19.8% and median PFS of 6.9 months (18, 19). The results in this study also compare favorably with those from phase III studies of regorafenib versus placebo (CORRECT and CONCUR), which were conducted in similar patient populations (40%–50% with four or more prior treatment regimens). In CORRECT, the median PFS with regorafenib was 1.9 months, with an ORR of 1.0%, and in CONCUR, the median PFS was 3.2 months, with an ORR of 4% (20, 21).

It should be noted that the MTD and dosing schedule of FTD/TPI in this study (25 mg/m² twice daily on days 1–5 of a 14-day cycle) differ from the approved monotherapy dose and schedule (35 mg/m² twice daily on days 1–5 and 8–12 of a 28-day cycle; ref. 10). The dose and schedule for the triplet combination was selected on the basis of earlier preclinical studies and the results of an initial phase I Japanese study, which evaluated the FTD/TPI–irinotecan combination in 10 patients with mCRC refractory to 5-FU and oxaliplatin (11–13). In the Japanese study, patients were administered 40–70 mg/m² twice daily FTD/TPI (on days 1–5 and 8–12 of a 28-day cycle) and a fixed dose of irinotecan (150 mg/m² on days 1 and 15). However, the standard FTD/TPI dosing schedule resulted in delayed myelotoxicity in that study: although preliminary activity was observed, all patients experienced grade ≥3 hematologic AEs (13). On this basis, lower starting doses of FTD/TPI were used in this study, and the FTD/TPI dosing schedule was adjusted to accommodate irinotecan dosing using a biweekly schedule, which maintains the same dose intensity as the standard regimen (28-day cycle). At the recommended dose in the expansion cohort, lower frequencies of grade ≥3 hematologic AEs were observed, supporting the current dose and dosing schedule.

The results of this study were also consistent with recent phase I reports of other FTD/TPI combinations in mCRC, although FTD/TPI was dosed at a lower level in the triplet combination than in the other combination regimens (22–25). In a phase I/II study of FTD/TPI plus bevacizumab in refractory mCRC, the recommended phase II dose was FTD/TPI 35 mg/m² (twice daily on days 1–5 and 8–12 of a 28-day cycle) and bevacizumab 5 mg/kg every 2 weeks (24). Three dose-escalation studies of FTD/TPI and oxaliplatin in patients with previously treated mCRC arrived at a MTD of FTD/TPI 35 mg/m² (twice daily on days 1–5 of a 14-day cycle) plus oxaliplatin 85 mg/m² (on day 1 of a 14-day cycle; refs. 22, 23, 25). These studies showed manageable safety and preliminary antitumor activity of the FTD/TPI-containing regimens in patients with previously treated mCRC (22–25).

The main limitation of this study was the nature of the analysis: this was a nonrandomized study with a small sample size. As only 24 patients were enrolled in the dose-expansion phase, no formal statistical analysis was performed to assess efficacy. In addition, as part of the study design, nearly all patients enrolled in the study had been previously treated with and had tolerated irinotecan. It is unclear whether similar efficacy/tolerability would be observed in a patient population with a higher incidence of irinotecan resistance or without prior exposure; although, it is likely that better responses may be observed in patients without prior exposure to irinotecan.

In conclusion, the results of this dose-escalation/expansion study indicate that the triplet combination of FTD/TPI, irinotecan, and bevacizumab is feasible in patients with mCRC refractory to standard therapies and that further evaluation of this regimen is warranted in this patient population. Given the standard use of bevacizumab–fluoropyrimidine–irinotecan chemotherapy combinations in first-line mCRC, these results may also support future exploration of the FTD/TPI–bevacizumab–irinotecan combination in earlier lines of treatment (2, 3, 14).

A.M. Varghese is an employee/paid consultant for Roche, and reports receiving other commercial research support from Illumina, Lilly, Bristol-Myers Squibb, and Silenseed. A.B. Benson is an employee/paid consultant for Envision, Guardant, Dava Onc, Bayer, LSK, Therabionic, Terumo, Lexicon, Incyte, ACCC, and Bristol-Myers Squibb, and reports receiving commercial research grants from Acerta, Celegene, Advanced Accelerator Applications, Novartis, Infinity Pharmaceuticals, Merck Sharp and Dohme, Taiho Pharmaceuticals, Bristol-Myers Squibb, Medimmune/AstraZeneca, Xencor, PreECOG, Astellas, Amgen, ECOG-ACRIN, and SynCore. H.S. Hochster is an employee/paid consultant for Taiho, Bayer, and Genentech. L. Makris and K. Hamada are employees/paid consultants for Taiho Oncology. J.D. Berlin is an employee/paid consultant for LSK biopharma and Bayer, and reports receiving commercial research grants from Taiho and Bayer. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.M. Varghese, A.B. Benson, L. Makris, K. Hamada, L.B. Saltz

Development of methodology: A.M. Varghese, L. Makris, K. Hamada, L.B. Saltz

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.M. Varghese, D.B. Cardin, J. Hersch, A.B. Benson, H.S. Hochster, J.D. Berlin, L.B. Saltz

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.M. Varghese, H.S. Hochster, L. Makris, K. Hamada, J.D. Berlin, L.B. Saltz

Writing, review, and/or revision of the manuscript: A.M. Varghese, D.B. Cardin, J. Hersch, A.B. Benson, H.S. Hochster, L. Makris, K. Hamada, J.D. Berlin, L.B. Saltz

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Hersch, L. Makris, K. Hamada, J.D. Berlin

Study supervision: A.M. Varghese, J.D. Berlin, L.B. Saltz

This study was sponsored by Taiho Oncology, Inc. and Taiho Pharmaceutical Co., Ltd. Medical writing and editorial assistance were provided by Vasupradha Vethantham, PhD and Lilly Ostrovsky of Scientific Connexions and funded by Taiho Oncology, Inc.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.